With the parity data, if some parts of your original data are missing, it is possible to resolve them. The parity data is like the check digit of your credit card number. This is because, in a RAID-5 array, the parity data of the amount equal to the size of a disk will be calculated when you write data into the array. You can replace the failed disk and rebuild the data in it by using the data in other N-1 disks. When you use N disks to build a RAID-5 array (or a raidz1 VDEV), you allow any one-disk failure to happen at one time. However, do I really need to sacrifice so much space for redundancy? If you what to have larger storage space with the same amount of disks, you may take a look at the solution realized by using parity data. If the budget is okay, you should go for it. This is the simplest way that requires minimal computation power for redundancy. The other half amount of the disks will be used to store a full copy of your data. In the case of mirroring (or RAID-1), you can only use half amount of your disks for data storage. They either simply mirror the data to a second disk or store some parity data in several extra disks. When it comes to redundancy, the general concept behind RAID and ZFS are quite similar. A software-RAID can be built by the mdadm command, while the ZFS-setup can be managed by zpool and zfs commands. I’d recommend using RAID or ZFS to set up such a large storage volume for home-usage. Also, we would require some redundancy such that the data in the large pool can survive under some hard disk failures. In a NAS, we usually combine many hard disks into a single large storage volume for convenience. Today, I’m going to take a note about how I build my NAS, which also works as a Time Machine backup place for my Macbook. However, these PCs may still be suitable for serving as a NAS server. It’s pretty common to have some PCs whose CPUs are no longer fast enough for our heavy tasks like rendering 8k-video or running numerical simulations. Building NAS with ZFS, AFP/Samba for Time Machine
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |